A light-emitting diode (LED) is a widely used semiconductor device whose main characteristic is that it will emit energy in the form of light, usually within the visible spectrum, when a current flows through the device. The basic mechanisms by which light-emitting diodes operate are well understood in this art and are set forth, for example, by Sze, PHYSICS OF SEMICONDUCTOR DEVICES, 2d Edition (1981) at pages 681-703. Silicon carbide-based LEDs are described in U.S. Pat. Nos. 4,918,497 and 5,027,168, both of which are assigned to the assignee of the present invention and incorporated entirely herein by reference.
As is well known to those familiar with semiconductor devices, light-emitting diodes, and the interactions between light and matter, the wavelength of light (i.e., its color) that can be emitted by a given semiconductor material is limited by the physical characteristics of that material, specifically its bandgap. The bandgap in a semiconductor material represents the amount of energy that separates a lower energy valence band and a higher energy conduction band in which carriers (electrons or holes) can reside in accordance with well-known principles of quantum mechanics. When electrons and holes travel across the bandgap and recombine, they will, under certain circumstances, emit energy in the form of light. Biasing a semiconductor p-n junction to produce a current flow is one way to obtain such recombinations and the visible light they emit. Because the wavelength of light is inversely proportional to its frequency, and its frequency is directly proportional to the corresponding energy transition, certain wavelengths of light cannot be obtained in materials that have relatively narrow bandgaps. For example, blue light is generally considered to be that visible light which is emitted in the 400-500 nanometer (nm) portion of the visible spectrum. It will be understood that 400-500 nm is a somewhat arbitrary range, and that wavelengths close to 400 nm are also considered to be violet, and those close to 500 nm to be green. Such wavelengths require energy transitions of at least 2.6 electron volts (eV) which means that light-emitting diodes that will emit blue light must be formed of materials that have a bandgap of at least 2.6 eV. Such materials include, in certain circumstances, zinc selenide (ZeSe), Group III nitrides (e.g. GaN, AlGaN, InGaN), diamond, (C) and silicon carbide (SiC).
Silicon carbide has a number of attractive features from an electronic standpoint. It has a high saturated electron-drift velocity, a wide bandgap, a high thermal conductivity, a high breakdown electric field, and is “hard” to radiation. The desirable theoretical characteristics of silicon carbide, and its potential as a source material for blue LEDs, have been well understood for a number of decades, dating back almost to the beginning of the semiconductor era. Nevertheless, the difficulties of working with silicon carbide have precluded most researchers from producing commercially successful devices from it.
For example, silicon carbide can crystallize in over 150 polytypes, many of which are separated by very small thermodynamic differences. As a result, and as well known to those familiar with crystal growth techniques of semiconductors and other materials, obtaining the necessary pure single crystals of silicon carbide, and the typical epitaxial or implanted layers that are generally desired or required in many semiconductor device structures, has long been a difficult task.
In recent years, however, the assignees of the present invention have made significant progress in surmounting the process difficulties presented by silicon carbide and in taking advantage of its desirable characteristics. These include success in the areas of sublimation growth of single crystals (e.g., U.S. Pat. No. 4,866,005 and its reissue Re34,861); growth of epitaxial layers of silicon carbide on single crystals (U.S. Pat. Nos. 4,912,063 and 4,912,064); implantation and activation of dopants into silicon carbide (U.S. Pat. No. 5,087,576); and etching techniques for silicon carbide (U.S. Pat. Nos. 4,865,685 and 4,981,551).
Building upon these successes, the assignees of the present invention have produced the first commercially viable blue light-emitting diodes in significant commercial quantities at reasonable prices. Such LEDs are thoroughly described in U.S. Pat. Nos. 4,918,497 and 5,027,168.
Silicon carbide, however, is an “indirect” semiconductor, meaning that when a radiative recombination occurs in SiC, some of the energy is released as a phonon rather than a photon, thus reducing the overall efficiency of the process. A representative discussion of “Luminescent Efficiency” is set forth in Sze, supra at §12.22 beginning on page 686.
Accordingly, over the last decade, interest has increased in the wide bandgap direct emitters, particularly the Group III nitrides. For example, gallium nitride (GaN) has a direct bandgap energy of 3.36 eV at room temperature (300K). Furthermore, by including other Group III elements, particularly aluminum (Al) and indium (In) in ternary and quaternary compounds, Group III nitrides can be tailored to a great degree to meet desired criteria of wavelength, conductivity, lattice matching, refractive index, and chemical stability. Exemplary (but not limiting) patents include U.S. Pat. Nos. 5,393,993, 5,523,589, 6,201,262 and 6,187,606, each of which are assigned to the present assignee and are incorporated entirely herein by reference. Exemplary pending applications include Ser. Nos. 09/154,363 filed Sep. 16, 1998 and 09/477,982 filed Jan. 5, 2000, both for “Vertical Geometry InGaN LED.” The contents of these applications are likewise incorporated entirely herein by reference.
For the time being, however, Group III nitride materials are not commonly available in bulk or substrate form. Instead, typical Group III nitride devices generally incorporate epitaxial layers of the nitrides on some other substrate material.
Sapphire (Al2O3) has been widely used as a substrate material for nitride devices. Sapphire offers optical transparency, chemical stability, and a manageable difference in lattice constant from most Group III nitrides. Sapphire cannot be conductively doped, however, and thus cannot form the basis of “vertical” devices, i.e. those in which ohmic contacts can be conveniently placed at opposite ends of the device.
Accordingly, the progress made in developing silicon carbide materials has also benefited the development of Group III nitride devices because SiC offers a better lattice match with most Group III nitrides than does sapphire and, perhaps most importantly, can be conductively doped. Thus, preferred Group III nitride device structures now incorporate vertical geometry using conductive SiC substrates. The patents referred to above incorporate these features.
Because an LED typically includes a diode structure (i.e., a p-n junction), commercial Group III nitride LEDs formed on a silicon carbide substrate generally include an n-type substrate and terminate in a p-type epitaxial layer, or alternatively, incorporate a p-type substrate and terminate in an n-type layer. The characteristics of silicon carbide, however, are such that the n-type of silicon carbide is somewhat easier to dope, and is more transparent when doped. Additionally, n-type semiconductors generally have a greater conductivity than do p-type semiconductors. Accordingly, the use of n-type layers wherever possible affords greater electrical conductivity (lower resistance) and optical transparency with resulting increases in light emission, efficiency, and current spreading for LED structures made therefrom.
Furthermore, producing successful ohmic contacts to p-type Group III nitride layers has often required a relatively high temperature anneal of the layer and the contact. For example, published European Patent Application 0 622 858 A2 to Nakamura et al. suggests that ohmic contacts to p-type layers require annealing at temperatures of at least 400° C., but recognizes that if the annealing temperature is too high, the Group III nitride compounds will begin to dissociate. Thus, the formation of ohmic contacts to p-type layers of Group III nitrides presents a compromise between the high temperatures desired for the anneal and the lower temperatures necessary to avoid dissociation or other degradation of the Group III materials.
Although it would be advantageous for a light emitting device to incorporate both an n-type substrate and an n-type Group III-nitride top layer, the presence of a p-n junction (hence a p-type layer) between two n-type layers would necessarily result in an n-p-n structure. As is known to those of skill in the art, a p-n junction rectifies current. That is, it permits net current to flow in only one direction (namely, from the p-type portion to the n-type portion). Thus, an n-p-n structure would prevent current flow in either the forward or reverse direction, thereby rendering the device inoperative.
Although these issues have been addressed to some extent in silicon carbide devices (e.g. U.S. Pat. No. 5,338,944; commonly assigned herewith and incorporated entirely herein by reference), they have not yet been addressed for the Group III-nitride devices. U.S. Pat. No. 5,338,944 discloses a light emitting diode formed on an n-type silicon carbide substrate with a degenerate junction structure for coupling the active layer to an n-type top layer while preventing n-p-n behavior between the n-type top layer, the active layer and the substrate. However, the device described in the '944 patent has a silicon carbide active region, which, as is discussed above, has a low efficiency of emission due to its indirect bandgap. In addition, Moreover, the device described in the '944 patent requires the formation of an ohmic contact to n-type silicon carbide after epitaxial deposition of the active layer, which typically requires additional dopant implantation and/or annealing steps.